Re: [linux-pm] PM: cannot hibernate -- BUG at kernel/workqueue.c:3659

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Tuesday, January 24, 2012, Jiri Slaby wrote:
> On 01/24/2012 11:36 PM, Rafael J. Wysocki wrote:
> > On Tuesday, January 24, 2012, Jiri Slaby wrote:
> >> On 01/24/2012 05:18 PM, Srivatsa S. Bhat wrote:
> >>> Hi Jiri,
> >>>
> >>> On 01/24/2012 08:35 PM, Jiri Slaby wrote:
> >>>
> >>>> Hi,
> >>>>
> >>>> this is a freshly booted system. When I do s2dsk, I see:
> >>>> ...
> >>>> Freezing remaining freezable tasks ... BUG: 'workqueue_freezing' is true!
> >>>> ------------[ cut here ]------------
> >>>> kernel BUG at /l/latest/linux/kernel/workqueue.c:3659!
> >>>> invalid opcode: 0000 [#1] SMP
> >>>> CPU 0
> >>>> Modules linked in:
> >>>>
> >>>> Pid: 2669, comm: s2disk Not tainted 3.3.0-rc1-next-20120124_64+ #1627
> >>>> Bochs Bochs
> >>>> RIP: 0010:[<ffffffff8107e365>]  [<ffffffff8107e365>]
> >>>> freeze_workqueues_begin+0x195/0x1a0
> >>>> RSP: 0018:ffff880046f01d68  EFLAGS: 00010292
> >>>> RAX: 0000000000000023 RBX: 0000000000000001 RCX: 00000000000000c9
> >>>> RDX: 0000000000000077 RSI: 0000000000000046 RDI: ffffffff81b51f7c
> >>>> RBP: ffff880046f01d98 R08: ffffffff81a9d760 R09: 0000000000000000
> >>>> R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
> >>>> R13: 00007fff579464dc R14: ffffffffffffffff R15: 0000000000000004
> >>>> FS:  00007f3c65d54700(0000) GS:ffff880049600000(0000) knlGS:0000000000000000
> >>>> CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
> >>>> CR2: 00007f3c64f58c20 CR3: 0000000045b64000 CR4: 00000000000006f0
> >>>> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> >>>> DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
> >>>> Process s2disk (pid: 2669, threadinfo ffff880046f00000, task
> >>>> ffff880047251980)
> >>>> Stack:
> >>>>  ffff880046f01d98 0000000000000001 0000000000000000 00007fff579464dc
> >>>>  ffffffffffffffff 0000000000000004 ffff880046f01e18 ffffffff81096cb9
> >>>>  00000000ffff0124 0000000000000004 ffff880046f01e18 000000004f1ec7d1
> >>>> Call Trace:
> >>>>  [<ffffffff81096cb9>] try_to_freeze_tasks+0x1b9/0x2d0
> >>>>  [<ffffffff81096ed5>] freeze_kernel_threads+0x25/0x90
> >>>>  [<ffffffff81097b55>] hibernation_snapshot+0x75/0x2e0
> >>>>  [<ffffffff8109d724>] snapshot_ioctl+0x314/0x4e0
> >>>>  [<ffffffff81130856>] do_vfs_ioctl+0x96/0x550
> >>>>  [<ffffffff8111ff7b>] ? vfs_write+0x10b/0x180
> >>>>  [<ffffffff81130d5a>] sys_ioctl+0x4a/0x80
> >>>>  [<ffffffff81630e22>] system_call_fastpath+0x16/0x1b
> >>>> Code: c7 c6 0a a4 92 81 48 c7 c7 16 65 92 81 31 c0 e8 19 94 5a 00 0f 0b
> >>>> 48 c7 c6 27 a4 92 81 48 c7 c7 16 65 92 81 31 c0 e8 02 94 5a 00 <0f> 0b
> >>>> 66 0f 1f 84 00 00 00 00 00 55 48 c7 c7 82 4b b9 81 48 89
> >>>> RIP  [<ffffffff8107e365>] freeze_workqueues_begin+0x195/0x1a0
> >>>>  RSP <ffff880046f01d68>
> >>>> ---[ end trace 632574abdc098963 ]---
> >>>>
> >>>
> >>>
> >>> I couldn't find any obvious root-cause from a quick check. Is this completely
> >>> reproducible upon a fresh boot?
> >>
> >> True.
> >>
> >> The cause is that the function is called twice:
> > 
> > Which function?
> 
> The one where the BUG is. Maybe the functions which should clear the
> flag is not called in between? See:
> 
> >>  [<ffffffff8107e206>] freeze_workqueues_begin+0x36/0x1b0
>                          ^^^^^^^^^^^^^^^^^^^^^^^
> >>  [<ffffffff81096cc9>] try_to_freeze_tasks+0x1b9/0x2d0
> >>  [<ffffffff81096ee5>] freeze_kernel_threads+0x25/0x90
> >>  [<ffffffff81097b65>] hibernation_snapshot+0x75/0x2e0
> >>  [<ffffffff8109d734>] snapshot_ioctl+0x314/0x4e0
> >>  [<ffffffff81130866>] do_vfs_ioctl+0x96/0x550
> >>  [<ffffffff8111ff8b>] ? vfs_write+0x10b/0x180
> >>  [<ffffffff81130d6a>] sys_ioctl+0x4a/0x80
> >>  [<ffffffff81630e22>] system_call_fastpath+0x16/0x1b
> >> (elapsed 0.03 seconds) done.
> ...
> >> Freezing remaining freezable tasks ... BUG: 'workqueue_freezing' is true!
> >> ------------[ cut here ]------------
> >> kernel BUG at /l/latest/linux/kernel/workqueue.c:3659!
> ...
> >> RIP: 0010:[<ffffffff8107e371>]  [<ffffffff8107e371>
> >> freeze_workqueues_begin+0x1a1/0x1b0
>    ^^^^^^^^^^^^^^^^^^^^^^^
> >> Call Trace:
> >>  [<ffffffff81096cc9>] try_to_freeze_tasks+0x1b9/0x2d0
> >>  [<ffffffff81096ee5>] freeze_kernel_threads+0x25/0x90
> >>  [<ffffffff81097b65>] hibernation_snapshot+0x75/0x2e0
> >>  [<ffffffff8109d734>] snapshot_ioctl+0x314/0x4e0
> >>  [<ffffffff81130866>] do_vfs_ioctl+0x96/0x550
> >>  [<ffffffff8111ff8b>] ? vfs_write+0x10b/0x180
> >>  [<ffffffff81130d6a>] sys_ioctl+0x4a/0x80
> >>  [<ffffffff81630e22>] system_call_fastpath+0x16/0x1b

Ah.  So this is linux-next, right?

Can you please test the linux-next branch of the linux-pm tree and see if
the problem is reproducible in there?

Rafael